Migration In-Flow and Out-Flow

Our department is always tasked with where residents are moving to and moving from. Whether its based by state or county. Is there information on migration based on age breakout (18-24, 25-34, 35-44, etc.) available down to the demographic level?
  • Hi Robin,

    I've uploaded two tabulations for you – they will need to be merged, as I'll explain (if this is what you are looking for).
    They contain:
    Weighted Person Count (Pwgtp)
    Unweighted Person Count (People)
    Person Level Household Count (PL_Wgtp)
    LoVal & HiVal (bracketing Pwgtp, also known as MOE)

    Broken out by:
    State of Residence (Pst, for Person Level State)
    Place from which they moved (either Migsp05 or Migsp12, depending on the file)
    Age (Age_v_RM_1)

    The file they are taken from is the 2009-2013 PUMS file, and in 2012 the census bureau changed the way they define places of migration. So there is one definition for the data collected in 2009-2011 and another for 2012-2013.

    To merge them, put any data that you want to merge side by side (e.g. you might not be interested in the two people that moved from Poland to Arizona in 2009-2011), matching up Pst, Migsp05/Migsp12, and Age. The add the two Pwgtp columns, the two People columns, and the two PL_Wgtp columns. It is not legitimate to add the LoVal and HiVal columns. So to get the total number of 18-24 year olds that moved from Texas to Nebraska, add 369 (Pwgt from line 13,241 of StMigsp05Age.csv) to 492 (Pwgt from line 12,075 of StMigsp12.csv) for a total of 861. If you want total unweighted person counts and household counts, you can add those also. (Note that it is not legitimate to add household counts across age groups – because some households that moved will have e.g. both 0-17 year olds and 18-24 year olds, those households would then be double counted).

    Something neat that I didn't expect when I ran this is that they also show in-state moves – e.g. people that moved from one place in Alabama to another place in Alabama.

    Everybody in the file shows up in both tabulations – if they are in the “Code classification is Not Applicable ...” category in one tab, they show up in the migration breakout of the other tab and vice versa.

    If you have any questions about this, (or if you need another dimension or volume), please ask,

    - John Grumbine
  • Hi Robin,
    Just to provide some more background, the Census Bureau produces a matrix of state-to-state migration flows based on ACS data:

    But they don't publish these data by age, which is why you need to use the Public Use Microdata Sample (PUMS) data to access that information.

  • In reply to John Grumbine:

    Do you know where I can find where people are coming from that live outside the United States?
  • In reply to Robin DiSalvo:

    Assuming that you mean people who have migrated to the US from other countries, and you want to know what those countries are, yes. Do you want to specify Hillsborough county? Any age breakouts? This can be done with the PUMS, I can do it with MAST or Gateway To MAST, or you could do it yourself with Gateway To MAST.
  • In reply to John Grumbine:

    Hi John,

    Yes, that's exactly what I am looking for - people who have migrated to the US from other countries. I am not familiar with PUMS or Gateway To MAST. Actually I need it for the Tampa-St. Petersburg-Clearwater MSA and Hillsborough County. No age breakouts needed - just the total number by country. I found table B05006: Place of Birth for the Foreign Born Population in the United States 2012-2016 in ACS but was not sure if this is what I am looking for.
  • In reply to Robin DiSalvo:

    I'm not sure PUMS is necessary. You might be able to get what you need from ACS table B05006, "PLACE OF BIRTH FOR THE FOREIGN-BORN POPULATION IN THE UNITED STATES": factfinder.census.gov/.../B05006

    This data is available down to the census tract level.
  • In reply to Bernie:

    Bernie - that's exactly the table I pulled - so I was on the right track afterall. Thanks!
  • In reply to Robin DiSalvo:

    I think that you are probably better off with that table than with the PUMS data that I can give you. The table is more current - the data that I'm working with is from 2014.
  • In reply to John Grumbine:

    Thanks for the confirmation John! You both have been a huge help to me.
  • In reply to Robin DiSalvo:

    This is totally unrelated, but have you looked at IRS County-to-County migration? It shows you not only where people are coming from (and leaving to), but also how much income they're bringing and taking with them. www.irs.gov/.../soi-tax-stats-migration-data
  • In reply to Bernie:

    No I haven't - but this will definitely come in handy. THANK YOU SO MUCH!!!!
  • In reply to Bernie:

    Following up on the table that added, you can use the Add/Remove Geographies tool to select the MSA's and counties that you're after by clicking through the drop down options. I went ahead and did that for the Table B05006. factfinder.census.gov/.../0500000US12057|320M300US1245300

    Two caveats to keep in mind. Census doesn't recommend comparing overlapping period estimates, so 2012-2016 estimates can be compared to 2007-2011 estimates, but shouldn't be compared to say, 2013-2017 estimates when the come out around December of this year.

    Second, some of the margins of error are very large relative to the estimates in this table, sometimes larger than the estimate itself. I would recommend taking the MOEs into account one way or another.